--On Wednesday, May 28, 2003 2:12 PM -0700 Kelsey Cummings <[EMAIL PROTECTED]> wrote:
On Wed, May 28, 2003 at 01:25:49AM -0700, Kelsey Cummings wrote:On Tue, May 27, 2003 at 09:21:34AM -0400, Adam Denenberg wrote: > I have seen this pop up before but with no real answer. If you have 2 > relay servers with equal MX costs sitting on the perimiter running SA, > how do you have it so they both share and update an "in-sync" db so > that both relay servers have up to date Bayes DBs? > > I remember reading that nfs wont work b/c of the flock() call, so > there must be another means no?
To my knowledge (I haven't looked at the code) the Bayes DBs use the same locking code as the AWL which is specificaly NFS safe. So, either rewrite it for SQL (and let us all know) or use NFS like we do.
Just a thought, why not run SA on the delivery host instead of the MX hosts? (I'm assume that both of your MX servers must forward the mail to a central server.)
-- Kelsey Cummings - [EMAIL PROTECTED] sonic.net
Your suggestion is the exact opposite of the intended function. The idea is to run two relays to share the load, or run as primary/secondary (MX), filter via SA, then deliver to the main mail server. You are proposing to dump the work back onto an already-overloaded main server, and remove the two "helpers" doing the filtering. Doesn't make sense to me. YMMV. --Alan
------------------------------------------------------- This SF.net email is sponsored by: eBay Get office equipment for less on eBay! http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5 _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
